Mismatch in impedance vs frequency curve (datasheet versus measurement)

[henryTheCat]

Hi @all!
I am new to the forum and wanted to ask a question about the impedance curve of a speaker.

I am rather new to the world of audio speakers and currently playing around to get a better grasp on the theory and practise of those components. I have borrowed a DATSV3 measurement kit and some speaker from a friend of mine to do so.

So as a first check (the big goal in the long run will be to couple the speaker in a box with an amp, but still a long way to go), I just took the plain speaker (not in any enclosure, but in free air) which is a scanspeak 18W-8542-10, and made an impedance versus frequency measurement with the DATS_V3 (the DATS was calibrated before).
Then I plotted the resulting curve against the curve in the datasheet. There is a mismatch, which I would call "big", but to be honest, I do not have a feeling how well a simple air measurement can be compared to the datasheet. What I can say is that the measurement itself is reproducable and does not vary too much if i hold the speaker in some different orientation, so there should be no big systematic error just because of "wrong" speaker positioning.

You see the two curves below. The resonance frequency is about 10Hz (!) shifted and the measured curve clearly shows a broader resonace (smaller quality factor?). Can someone explain why this could be the case?

Is this something to be expected because of the described measurement setup? If yes, is there something I can/should do to improve the measurement?

Thanks a lot!

 

[D Murphy]

Hi @all!
I am new to the forum and wanted to ask a question about the impedance curve of a speaker.

I am rather new to the world of audio speakers and currently playing around to get a better grasp on the theory and practise of those components. I have borrowed a DATSV3 measurement kit and some speaker from a friend of mine to do so.

So as a first check (the big goal in the long run will be to couple the speaker in a box with an amp, but still a long way to go), I just took the plain speaker (not in any enclosure, but in free air) which is a scanspeak 18W-8542-10, and made an impedance versus frequency measurement with the DATS_V3 (the DATS was calibrated before).
Then I plotted the resulting curve against the curve in the datasheet. There is a mismatch, which I would call "big", but to be honest, I do not have a feeling how well a simple air measurement can be compared to the datasheet. What I can say is that the measurement itself is reproducable and does not vary too much if i hold the speaker in some different orientation, so there should be no big systematic error just because of "wrong" speaker positioning.

You see the two curves below. The resonance frequency is about 10Hz (!) shifted and the measured curve clearly shows a broader resonace (smaller quality factor?). Can someone explain why this could be the case?

Is this something to be expected because of the described measurement setup? If yes, is there something I can/should do to improve the measurement?

Thanks a lot!
View attachment 43071

That's a head scratcher unless you've reversed the labels. Virtually every Scan woofer I've measured has had a higher Fs than specified, even after I've broken it in extensively. Did you take the measurement immediately after playing the woofer? If so, try letting it cool down first. I have DATS 3, and it's accurate for simple impedance measurements like this. (But like all the earlier versions, it doesn't do as well on inductance measurements.)

 

[Speedskater]

Try running a curve on some power resistors of known value. They should be straight lines.

 

[henryTheCat]

That's a head scratcher unless you've reversed the labels. Virtually every Scan woofer I've measure has had a higher Fs than specified, even after I've broken it in extensively. Did you take the measurement immediately after playing the woofer? If so, try letting it cool down first. I have DATS 3, and it's accurate for simple impedance measurements like this. (But like all the earlier versions, it doesn't do as well on inductance measurements.)

I have checked the labels and I cannot see an error in that respect. The speaker is rather new and it was at room temperature, only a few measurements were done with the DATS, no previous playback or warming.

What´s also interesting:
If I plot the impedance curve by just using the basic spice model (using the values Re, Le, Fs, Qms and Qes only) the curve fits more or less well to the measured DATS curve (of course not for high frequency because semi-inductance is not modelled correctly, but the resonance peak fits well).

I also see that the T/S datasheet values used for spice simulation fit well to the T/S values determined by DATS free-air, also the resonance frequency is only 3Hz higher in the measurement.

	Re	Le	Fs	Qms	Qes
DATS	5.9	0.14m	45.8	1.74	0.47
Datasheet	5.7	0.2m	43	1.7	0.44

So: T/S parameters agree between measurement and datasheet (beside some small shift of ~3Hz in fs), therefore also the spice model agrees well with the measurement , not much difference whether the T/S from Datasheet or T/S from measurement are used

.....but plotting a impedance curve based on those values does not produce the graph in the datasheet (fs about ~10Hz higher).

The only explanation I have for that is that the graph does not show a free-air measurement OR the T/S parameters are not good for free-air?

What do you think?

BR,
Franz

 

[henryTheCat]

Try running a curve on some power resistors of known value. They should be straight lines.

Yes, they are straight lines. I checked that and the measurement of the DATS itself seems ok (see also my post just above)